Classification of Raw Minced Beef, Chicken, and Pork Using AS7341 Spectrophotometer Sensor with Naive Bayes Method Humaidillah Kurniadi Wardana (a*), (c), Endarko (a), Agus Muhamad Hatta (b)
(a) Department of Physics, Institut Teknologi Sepuluh Nopember (ITS), Kampus ITS Sukolilo, Surabaya 60111, Indonesia
a* humaidillahwardana[at]unhasy.ac.id
(b) Department of Physics Engineering, Institut Teknologi Sepuluh Nopember (ITS), Kampus ITS Sukolilo, Surabaya 60111, Indonesia
(c) Electrical Engineering Studi Program, Universitas Hasyim Asy^ari Tebuireng, Cukir-Diwek, Jombag 61471, Indonesia
Abstract
Meat has a high nutritional source needed by the human body because it contains sources of protein, vitamins, minerals, amino acids and fatty acids. The rise of cases of adulteration of beef and chicken meat mixed with pork often occurs every year. For this reason, a standard of meat purity is needed so that it is halal and safe for consumption. This research proposes the manufacture of a portable spectrophotometer that is easy to use, low in price, and able to distinguish beef, chicken, pork. The spectrophotometer is made using an AS7341 sensor equipped with 1 LED light source, 11 spectral channels with a range of 400-940 nm, raspberry pi 4B as a microcontroller, and the results are displayed on the LCD. The results show that the system can distinguish beef, chicken, pork, beef-pork mixture with a ratio of (7:3, 5:5, 1:9) grams, and chicken-pork mixture with a ratio of (7:3, 5:5, 1:9) grams using the Naive Bayes Classifier method. It is obtained that the evaluation value of the beef, pork, and beef-pork mixture classification system with a total data set of 753 data, training data and testing data has a ratio of (70:30), (80:20) getting an accuracy value of 1.0, precision 1.0, recall 1.0, F1-score 1.0, and AUC 1.0 (excellent classification). While the evaluation of the classification system for chicken, pork, and chicken-pig mixtures with a total data set of 758 data, training data and testing data has a ratio of (70:30), (80:20) getting an accuracy value of 1.0, precision 1.0, recall 1.0, F1-score 1.0, and AUC of 1.0 (excellent classification).
